Incorporating the multiple linear regression with the neural network to the form design of product image

A consumers´ psychological perception (CPP) of a product is induced by its appearance, and thereby product form plays a vital role for the commercial success of a product. This study proposes an incorporated design approach combining a multiple linear regression technique with a back-propagation neural network to aid product designers incorporate CPPs of product forms in the design process. To demonstrate the feasibility of the incorporated approach, this study considers the design of an automobile profile and then performs a series of evaluation trials to establish the relationship between the automobile profile and the CPPs. The results of the evaluation trials are used to construct the MLRBPN models capable of predicting the likely CPP to any automobile profile designed in accordance with the numerical automobile profile definition. Although the automobile profile is chosen as an example, the concept of the proposed approach is equally applicable to other consumer product form.